Neural and behavioural indices of face processing in siblings of children with autism spectrum disorder (ASD): A longitudinal study from infancy to mid-childhood

Abstract

Impaired face processing is proposed to play a key role in the early development of autism spectrum disorder (ASD) and to be an endophenotypic trait which indexes genetic risk for the disorder. However, no published work has examined the development of face processing abilities from infancy into the school-age years and how they relate to ASD symptoms in individuals with or at high-risk for ASD. In this novel study we investigated neural and behavioural measures of face processing at age 7 months and again in mid-childhood (age 7 years) as well as social-communication and sensory symptoms in siblings at high (n = 42) and low (n = 35) familial risk for ASD. In mid-childhood, high-risk siblings showed atypical P1 and N170 event-related potential correlates of face processing and, for high-risk boys only, poorer face and object recognition ability compared to low-risk siblings. These neural and behavioural atypicalities were associated with each other and with higher social-communication and sensory symptoms in mid-childhood. Additionally, more atypical neural correlates of object (but not face) processing in infancy were associated with less right-lateralised (more atypical) N170 amplitudes and greater social-communication problems in mid-childhood. The implications for models of face processing in ASD are discussed.

Keywords: Autism spectrum disorder (ASD); Development; EEG; Face processing; Infant siblings.

Fig. 1

Stimuli used in the face recognition and face processing tasks in mid-childhood. Panel (A) shows examples of the face, car, body and scene stimuli used in the face recognition task in mid-childhood. Panel (B) shows examples of the upright and inverted face and fixation stimuli used in the EEG face processing task in mid-childhood.

Fig. 2

Performance in the face recognition task in mid-childhood. Boxplots display the group means (black line) and individual participants' scores (circles) for accuracy and RT performance in the face recognition task in mid-childhood. Panel (A) shows accuracy performance (% correct trials per condition) and Panel B shows the mean-of median RT for correctly recognised trials (ms). HR and LR group means and individual scores are presented in separate columns. Children in the LR group are indicated by grey circles; children in the HR group are indicated by blue circles, with the HR children who met diagnostic criteria for ASD highlighted in yellow.

Fig. 3

Grand average waveforms and topographical plots for the P1 component by group and condition. Panel (A) shows the grand average stimulus-locked waveforms displaying the P1 ERP component for upright and inverted faces by HR and LR group at electrode O1 (left hemisphere, top) and electrode O2 (right hemisphere, bottom). Black line = Grand average for the upright face condition in the LR group. Red line = Grand average for the inverted face condition in the LR group. Blue line = Grand average for the upright face condition in the HR group. Blue line = Grand average for the inverted face condition in the HR group. Panel (B) shows the topographical maps of the P1 component by group (LR, HR) and condition (upright and inverted faces).

Fig. 4

Grand average waveforms and topographical plots for the N170 component by group and condition. Panel (A) shows the grand average stimulus-locked waveforms displaying the N170 ERP component for upright and inverted faces by HR and LR group at electrode P7 (left hemisphere, top) and electrode P8 (right hemisphere, bottom). Black line = Grand average for the upright face condition in the LR group. Red line = Grand average for the inverted face condition in the LR group. Blue line = Grand average for the upright face condition in the HR group. Blue line = Grand average for the inverted face condition in the HR group. Panel (B) shows the topographical maps of the N170 component by group (LR, HR) and condition (upright and inverted faces).

Fig. 5

Cross-sectional associations between face recognition performance, face processing ERP indices and ASD symptoms in mid-childhood. Scatterplots show the associations between face recognition performance, ERP indices of face processing and ASD symptoms in mid-childhood in the HR group. The blue circles indicate data from the HR children without ASD and yellow circles indicate HR children with ASD; the regression lines represent the association between the variables in the HR group (HR-ASD and HR-non-ASD children combined). Black asterisks represent data points from the LR group and are shown only for visual comparison with the HR group associations. Panel (A) shows the negative association between RT for correctly recognising faces in the face recognition task and the extent to which the N170 ERP component for faces was lateralised to the right hemisphere; faster RTs were associated with greater right-lateralisation of the N170. Panel (B) shows the positive association between RTs for correctly recognising face stimuli in the face recognition task and SRS-2 scores; faster RTs were associated with fewer social-communication problems. Panel (C) shows the negative association between lateralisation of the N170 ERP component and SRS-2 scores; greater right-lateralisation of the N170 was associated with fewer social-communication problems. Panel (D) shows the positive association between lateralisation of the N170 ERP component and SSP scores; greater right-lateralisation of the N170 was associated with fewer sensory symptoms (higher SSP scores).

Fig. 6

Longitudinal associations between face processing ERP indices in infancy and mid-childhood and mid-childhood ASD symptoms. Scatterplots show the associations between ERP indices of face processing at age 7-months and in mid-childhood and ASD symptoms in mid-childhood. The blue circles indicate data from the HR children without ASD and yellow circles indicate HR children with ASD; the regression lines represent the association between the variables in the HR group (HR-ASD and HR-non-ASD children combined). Black asterisks represent data points from the LR group and are shown only for visual comparison with the HR group associations. Panel (A) shows the negative association between the N290 amplitude difference score for face versus noise stimuli in infancy and SRS-2 scores in mid-childhood; larger (more negative) N290 difference scores (indicating larger N290 amplitude for face vs noise stimuli) were associated with more severe social-communication problems in mid-childhood. Panel (B) shows the positive association between the N290 amplitude difference score in infancy and the extent to which the N170 was right-lateralised in mid-childhood; larger (more negative) N290 difference scores (larger amplitudes for faces vs noise) were associated with less right-lateralisation of the N170 in mid-childhood. Panel (C) shows the negative association between the N290 difference score in infancy and the extent to which latency of the P1 was slower for inverted than upright faces in mid-childhood; larger (more negative) N290 difference scores (larger amplitudes for faces vs noise) were associated with larger latency increases for inverted versus upright faces in mid-childhood.